diff options
| author | Davide Italiano <davide@freebsd.org> | 2015-11-29 20:58:04 +0000 |
|---|---|---|
| committer | Davide Italiano <davide@freebsd.org> | 2015-11-29 20:58:04 +0000 |
| commit | b8b7133c9421e9cd1af7c2a525e207cae5316f6f (patch) | |
| tree | 425b15fcd5af8de692970899cc730ac2f747e5a8 /llvm/lib/Target/WebAssembly/WebAssemblyFrameLowering.cpp | |
| parent | a5c0493ddbb059c54acdfe1004cd3b702a0a5c95 (diff) | |
| download | llvm-b8b7133c9421e9cd1af7c2a525e207cae5316f6f.zip llvm-b8b7133c9421e9cd1af7c2a525e207cae5316f6f.tar.gz llvm-b8b7133c9421e9cd1af7c2a525e207cae5316f6f.tar.bz2 | |
[SimplifyLibCalls] Tranform log(pow(x, y)) -> y*log(x).
This one is enabled only under -ffast-math. There are cases where the
difference between the value computed and the correct value is huge
even for ffast-math, e.g. as Steven pointed out:
x = -1, y = -4
log(pow(-1), 4) = 0
4*log(-1) = NaN
I checked what GCC does and apparently they do the same optimization
(which result in the dramatic difference). Future work might try to
make this (slightly) less worse.
Differential Revision: http://reviews.llvm.org/D14400
llvm-svn: 254263
Diffstat (limited to 'llvm/lib/Target/WebAssembly/WebAssemblyFrameLowering.cpp')
0 files changed, 0 insertions, 0 deletions